> Went on to do more "complex" things Have you considered moving in the other direction? There's lots of work -- and value -- in bog-standard software with no

> Went on to do more "complex" things Have you considered moving in the other direction? There's lots of work -- and value -- in bog-standard software with no mathematical/statistical/computational cleverness. Occasionally you can see an opportunity to fruitfully apply a mildly clever quantitative technique and see it actually get used -- but >95% of the time there's usually necessary grunt work to do. How about trying contract work on enterprise software projects? E.g. pick a huge boring-sounding enterprise that has some weird legacy* software systems built for internal use. If you find the right one, they will have (1) users who use the system every day to do their job, and (2) lashings of technical debt, and they may be quite excited to have anyone on board to push forward efforts in "unit testing" or "continuous integration" or "implementing exception handling correctly". inspiration: https://www.butternotes.com/blog/catfish-programmers https://www.butternotes.com/blog/catfish-programmers http://edw519.posthaven.com/willie-sutton-would-be-an-enterprise-programmer http://edw519.posthaven.com/willie-sutton-would-be-an-enterp... * legacy doesn't necessarily mean old -- maybe the system uses some trendy new tech stack, but no one wrote any tests, and the first wave of contractors already left